Summary of Machine Learning Techniques with Fairness For Prediction Of Completion Of Drug and Alcohol Rehabilitation, by Karen Roberts-licklider and Theodore Trafalis
Machine Learning Techniques with Fairness for Prediction of Completion of Drug and Alcohol Rehabilitation
by Karen Roberts-Licklider, Theodore Trafalis
First submitted to arxiv on: 23 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
Summary difficulty | Written by | Summary |
---|---|---|
High | Paper authors | High Difficulty Summary Read the original abstract here |
Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This study explores the use of machine learning models to predict whether individuals will complete a drug and alcohol rehabilitation program and their attendance frequency. The researchers utilize demographic data from Substance Abuse and Mental Health Services Administration (SAMHSA) for admissions and discharge data from Oklahoma-based centers. To address the categorical nature of this data, binary encoding is employed, along with fairness measures to mitigate bias in nine demographic variables. Various kernel methods, including linear, polynomial, sigmoid, and radial basis functions, are compared using support vector machines at different parameter ranges to determine optimal values. These results are then contrasted with decision trees, random forests, and neural networks. The study also employs Synthetic Minority Oversampling Technique Nominal (SMOTEN) for categorical data, imputation for missing data, and intersectionalization of the nine bias variables to mitigate bias. The paper explores various fairness metrics, including Disparate Impact, Statistical Parity difference, Conditional Statistical Parity Ratio, Demographic Parity, Demographic Parity Ratio, Equalized Odds, Equalized Odds Ratio, Equal Opportunity, and Equalized Opportunity Ratio, in both binary and multiclass scenarios. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary This study tries to figure out if people will finish a program that helps them stop drinking and using drugs. They also want to know how many times people go to these programs. To do this, they use information about people who have gone to these programs before. The data is special because it’s not just numbers, but also things like gender and age. So, the researchers use a special trick called binary encoding to make the data work better with computer models. They also want to make sure that the models are fair, so they test different ways to make sure the results aren’t biased against certain groups of people. The study compares different types of computer models, like decision trees and neural networks, to see which one works best. They also use a special technique called SMOTEN to balance the data and make it more accurate. |
Keywords
» Artificial intelligence » Machine learning » Sigmoid